How much has the planet warmed, and where?

Steps

1. GISTEMP v4: graph_data global annual series (No_Smoothing and NASA's Lowess(5) curve), ZonAnn.Ts+dSST.csv (zonal annual means, 14 columns) and gistemp1200_GHCNv4_ERSSTv5.nc (monthly 2 by 2 degree anomalies); the grid layer is the mean of the 12 months of 2025 per cell. NOAA GML co2_mm_mlo.csv (monthly mean and deseasonalised trend) and co2_annmean_mlo.csv.
2. Anomalies are differences from the 1951 to 1980 average of the same place and month, so a value is a change, not a temperature. The global series is NASA's; bands are NASA's zonal means; the 2 degree grid is used for the globe layer and checked against NASA's published means. The ten warmest years are ranked on the published values rounded to 0.01 degree.
3. The globe layer data/space/gistemp_grid_2025.json holds the 2025 annual mean per 2 by 2 degree cell (the mean of the 12 months where at least 6 have data), quantized to tenths of a degree, 45 KB.

146 yearly rows (1880 to 2025), 823 monthly carbon dioxide rows, 8 latitude bands, 16,200 grid cells. Unit: degrees Celsius against the 1951 to 1980 average (temperature anomaly); parts per million of dry air (carbon dioxide). Every derived figure on this page (changes, indexes, averages, counts) is computed from the published rows, and the table under each chart says how.

Checks

• NASA's zonal file has the same global value as the global annual series for every year, and the two hemispheres average to it; the mean of the global series over 1951 to 1980 is 0, as the base period requires.
• The eight latitude bands, weighted by area, reproduce the global value (mean difference 0.016 °C over 146 years; the largest, 0.102 °C, is in a year with sparse polar data).
• The 2 by 2 degree grid is checked before use: its area-weighted global mean for 2025 is 1.203 °C against NASA's published 1.18 °C (NASA fills missing areas differently), coverage 99.6% of the Earth's area, and each latitude band agrees with the zonal file within 0.1 °C.
• The monthly carbon dioxide file's yearly means equal NOAA's annual mean file in all 67 full years (largest difference 0.006 ppm); the monthly series has no gaps.
• The GISTEMP citation, the 1951 to 1980 base period, the uncertainty statement (about 0.05 °C after 1960), the pre-industrial adjustment (about 0.19 °C) and NOAA's data-use text are read from the NASA and NOAA pages by the script, which stops if their wording changes.

Caveats

• GISS puts the uncertainty of annual global means after 1960 at about 0.05 degrees Celsius, larger before. GISTEMP is one of several estimates (NOAA, the Hadley Centre and Berkeley Earth publish others) that agree on the shape and differ by hundredths of a degree; ranks of close years can differ between them. The 1951 to 1980 base is a convention: against the 1850 to 1900 pre-industrial period, GISS says to add about 0.19 degrees Celsius (as of January 2025). The polar regions before the 1950s are mostly unmeasured. The Mauna Loa record is one site, which tracks the global trend of this well-mixed gas; the January to April 2023 eruption gap was filled from the Maunakea site (NOAA file note).
• NASA's server data.giss.nasa.gov refused connections from Oct 10, 2026 (after the global series and zonal file had been read from it). The gridded file and the GISS home and FAQ pages were therefore read from the Internet Archive's copies of the same NASA URLs (the grid file captured Sep 6, 2026, NASA's file dated Aug 10, 2026, 25,836,095 bytes, GISTEMP through July 2026). The script checks the grid against NASA's own published 2025 means (global and by latitude band) before using it.
• An anomaly is a change from the 1951 to 1980 average, not a temperature; different base periods shift all values by a constant. Rankings of years within a few hundredths of a degree are not meaningful.

Releases and revisions

• GISTEMP is updated monthly and past values change slightly with each release; this is the file read on 2026-10-08. Re-run space_warming.py to refresh.
• NASA recomputes the whole record each month, so past values can move by a few hundredths of a degree.